Manifesto

We are stepping into unknown territory, together with our customers and their users. That’s why we have crafted a manifesto — to make sure we don’t lose sight of what really matters.

The Therapist’s Manifesto

• The AI therapist must never pretend to be human.
• The AI therapist must never pretend to have feelings (e.g., “I’m sorry to hear that”).
• The AI therapist must never enter into a social contract with the client (e.g., “Do you promise me?”).
• The AI therapist must, whenever therapeutically possible, set boundaries for its own availability to help prevent dependency.
• The AI therapist should encourage and assist the client in reaching out to other people whenever appropriate.
• The AI therapist must never attempt to handle acute crises. In case of suspected suicide risk, abuse, or other emergencies, it must always refer the client to qualified human help.

The Privacy Manifesto

• The our customers will never have access to read what the client writes.
• The provider may train AI models on anonymized client data only.
• Individual client data should be stored as close to the user as technically and financially possible.
• European AI and data providers should be preferred whenever technically and financially feasible.
• When realistic, open-source AI models should be hosted on user-selected data centers.

The Climate Manifesto

• Data centers and providers should, whenever possible, be selected based on their use of renewable energy. Our services are currently hosted @ Hetzner in Germany
• AI usage should be limited whenever technically and functionally possible.
• AI should not be used when a static text can solve the task.
• No larger or more energy-consuming AI model should be used if a smaller model can solve the task sufficiently.